Almost at the same time, several Bitcoin mining companies switched their operations from "mining" to "AI", as computing power and electricity began to serve another narrative: Hut 8 boldly signed a long-term lease agreement for an AI data center worth approximately $9.8 billion, while IREN secured a cloud service contract worth about $2.8 billion. These two AI-related collaborations, each exceeding a hundred billion, are seen as a signal of the mining companies’ complete shift towards data center infrastructure. Capital quickly reacted—shortly after the news broke, during Monday's early trading, stock prices of mining companies like IREN, Cipher Digital, CleanSpark, Hut 8, and MARA Holdings collectively surged, with gains reaching at least double digits, and the "mining companies + AI" label was swiftly marked up on the trading floor. However, on the other side, economist Peter Schiff provided a cold reminder with the AI-related ETF $SPCX: its current price is about 11% lower than its IPO issue price and has dropped over 46% from its historical high, now trading below $121. He warned that the AI stock bubble may have already burst and cautioned investors to be alert to the risk of valuation adjustments—the tension between the mining companies’ heavy bets on AI computing power, the short-term stock price frenzy, and the bubble alarm was thus drawn to extremes on the same timeline.
The Sudden Turnaround of Bitcoin Mining Plants into Computing Power Bases
On the same timeline where the bubble alarm has been repeatedly sounded, Bitcoin mining companies have long realized that relying solely on mining cannot support a long-term predictable cash flow statement. With block rewards halving every four years, electricity prices fluctuating with policy and energy cycles, and coin prices switching violently between bull and bear markets, these triple variables combine into a familiar yet cruel destiny: the number of Bitcoins produced is decreasing, while revenue is highly dependent on market conditions and cost windows. Thus, mining companies holding large-scale electricity and data center resources began to calculate another set of formulas—using the same electricity and space, if not just serving "hashing," but instead transitioning to hosting and operating AI computing power, could they upgrade a mining plant into a more broadly defined "digital infrastructure" base.
LM Funding America's actions can be considered a textbook case of this turnaround: the company directly changed its name to PowerCompute and repositioned itself as a provider of AI computing power infrastructure. It possesses approximately 26 megawatts of its own electricity, which in the context of traditional mining is merely considered "an advantage in electricity prices," but under the AI narrative is quickly reassessed as a foundational resource that can be transformed into GPU and server deployment capabilities. Mining companies no longer focus solely on “computing power mining,” but are trying to tell a composite business story that utilizes existing electricity, data centers, and operational systems to deliver diversified computing power services. Meanwhile, Strategy’s CEO Phong Le sent out a markedly different signal—his company did not perceive any market impact while selling approximately $200 million worth of Bitcoin. He emphasized, "We are not leaving," indicating that some participants choose to expand into AI while continuing to bet on Bitcoin, rather than completely abandoning their original cryptocurrency business. This has made the process of transforming mining plants from single mining sites into computing power bases more about adding a new narrative rather than hastily escaping the old world.
Hundred Billion Long-Term Contracts: Bet or Bind?
While Strategy is still expressing loyalty to Bitcoin, another frontline has quietly been delineated—Hut 8 signed a long-term lease agreement for an AI data center worth approximately $9.8 billion, while IREN obtained a cloud service contract worth around $2.8 billion. The amounts for these two contracts far exceed the typical scales traditional mining companies could secure for single mining businesses, and details regarding the contracting parties and specific lease terms have not been disclosed. The only certainty for the outside world is that they are categorized as AI-related infrastructure collaborations: on one side, the data centers, electricity, and cooling capabilities are packaged into "data centers," and on the other, computing power directly interfaces with AI demand in the form of cloud services, forcibly pulling mining companies from the “mining sites” betting on coin prices into a new narrative of “computing service providers.”
From a numerical perspective, this constitutes a high-stakes gamble on the future demand for AI computing power that seems almost irreversible. Long-term leases and cloud service contracts mean revenues will be settled gradually over the coming years, while costs—new racks, expanded electricity, maintenance of data centers—are also locked into the gears of time. Before the contracts have even completed a billing period, the capital market has already provided its immediate feedback: shortly after the relevant news was announced, during Monday's early trading, stock prices of IREN, Cipher Digital, CleanSpark, Hut 8, and MARA Holdings collectively surged, with an increase of at least 11%, as funds voted in favor of the “mining companies + AI” story. However, the short-term frenzy in stock prices is merely a rapid pricing of the narrative; whether these billion-dollar long-term commitments will become a moat that withstands cycles or transform into difficult-to-backtrack shackles amid future changes in AI demand and pricing systems remains uncertain.
AI Model Price Cuts: Another Side of Computing Power Demand
While mining companies were cheering for hundred-billion-level computing contracts, another side of the AI industry quietly initiated a "price-cutting competition." Low-cost large models like Kimi K3 burst onto the scene, advertised as providing reasoning and training services at lower prices and higher cost-effectiveness, directly targeting the traditional high-priced service providers that enjoyed premiums over the past two years. This is not a simple technological iteration but a rehearsal of malignant competition: model layer and application layer companies are forced to lower their quotes and extend their payback periods, and the previously documented high-margin, high-growth assumptions in pitch materials are now being questioned by the market. Economist Peter Schiff’s warning that "the AI stock bubble may have burst," along with the over 46% drop of $SPCX from its peak, actually reflects this systemic repricing of profit margins and expectations.
Once application layer companies experience compressed profits and cooler valuations, their demand for upstream computing power and data centers will no longer just increase. Executives signing large-scale long-term computing and cloud service contracts, like those with Hut 8 and IREN, will become more cautious, as the scale, term, and pricing terms of the contracts may be renegotiated. The problem is that most Bitcoin mining companies' AI transformation focuses on the infrastructure side—providing electricity, data centers, and computing power leasing—rather than directly participating in model research and terminal applications. Their perception of demand changes is often lagging: when the application end has already cut budgets and reduced incremental contracts, mining companies may just have completed their electricity expansions, cabinet constructions, and equipment procurements. On the surface, they seem to stand in the "selling shovels" position during the model explosion, but in reality, they may be pushed into an uncontrollable gamble on computing power due to the dual pressures of overexpansion and a reversal of contract bargaining power during the next cycle of demand and profit repricing.
Peter Schiff’s Bearish Outlook: AI Stocks Halved from Highs
At the same juncture when mining companies dove into the "big AI computing order" narrative, economist Peter Schiff, known for exposing asset bubbles, chose to stand on the opposite side. He has spent years continually questioning the high valuations of technology stocks and cryptocurrency assets, and now he is turning his criticism to the AI sector, bluntly stating, "The AI stock bubble may have already begun to burst." He believes that investors once gave AI companies overly idealized growth curves, which are now being corrected through prices. He provides a specific example—AI-related ETF or stock $SPCX: compared to its IPO price, it has fallen about 11%, and it is down over 46% from its historical peak, with a price below $121. In his narrative, this is not just a single asset's stall but a reflection of the entire sector’s regression from manic expectations to sober discounting.
Schiff’s logic is not complex: as low-cost models like Kimi K3 emerge, the pricing power of computing and algorithms begins to slip from the hands of a few top players, and valuations built on the premise of "high barriers, high margins, long-term monopolies" must be reexamined. He warns investors not to focus solely on the total contract value and revenue guidance but to ask—under the future scenario of intensified competition and declining service prices, do these AI companies still deserve their once high growth expectations and valuation premiums? For mining companies busy expanding data centers and signing ten-year computing contracts, this bearish outlook is not baseless but highlights a less than appealing premise: if AI stock prices are already on the path of bubble deflation, then the computing power infrastructure built around them must also confront the reality of valuation reductions following slowed growth stories.
The Crossroad of Mining Companies' Gamble on AI and Bubble Alarms
As Hut 8 and IREN package and sell their future ten years' worth of data centers and electricity to AI clients for approximately $9.8 billion and $2.8 billion, respectively, the subsequent stock price surge of at least 11% during Monday's early trading seems to cast a vote of approval for the "mining companies + AI" narrative; yet on the other side, low-cost models like Kimi K3 are pulling down the unit price of computing services, with $SPCX dropping over 46% from its peak, and Peter Schiff loudly reminding that the AI stock bubble may have already burst. These realities compel the market to reprice the hundred-billion-level contracts and heavy asset expansions that mining companies have just secured. If AI valuations continue to adjust, the long-term leases and cloud service contracts currently seen as “cycle hedges” may tomorrow become overpriced capacity bindings, requiring a reevaluation of profit models, leverage, and depreciation. Bitcoin price cycles, electricity costs, real demand for AI computing power, and the speed of valuation correction will together determine whether these contracts serve as a safety net or as leverage for amplifying negative impacts. Strategy's Phong Le’s assertion of "not leaving Bitcoin" reflects a dual exposure—betting simultaneously on the cryptocurrency economy and AI infrastructure, where mining companies either seize the benefits of the next round of foundational computing power, transforming volatility into recurring income, or endure systemic risks that are more concentrated and harder to hedge than traditional participants during the next round of tech bubble retraction.
Join our community to discuss and grow stronger together!
AiCoin exclusive Hyperliquid benefit: https://app.hyperliquid.xyz/join/AICOIN88
AiCoin exclusive Aster benefit: https://www.asterdex.com/zh-CN/referral/9C50e2
On-chain Telegram community: https://t.me/AiCoinWhaleData
On-chain community: https://www.aicoin.com/link/chat?cid=N6OVMor5g
AiCoin on-chain Twitter: https://x.com/aicoinwhaledata
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
